CN109830072A - The valuation of view-based access control model identification and cash device, control system and method - Google Patents

Abstract

The invention discloses a kind of valuation of view-based access control model identification and cash devices, control system and method, device includes host, customer service touch screen, customer's touch screen, vegetable identification camera, service plate placement plate and auxiliary multifunctional brush card machine, pressure sensor is set in the bottom surface of service plate placement plate, customer is respectively set above and below the display area of customer's touch screen and identifies camera and vegetable identification secondary light source.Control system includes vegetable recording module, Customer Information recording module and clearing payment module to realize that fast brushing face is paid.Method using efficiently and stability training pattern learn, it is ensured that vegetable typing, face typing and payment and settlement accuracy.The present invention solves the practical problems such as the custom queueing time is long in actual scene, settlement efficiency is low, vegetable data difficulty is retained, and a variety of client's actual demands such as improves customer experience, reduces queuing time, precisely analyzes data, cuts operating costs to realize.

Description

The valuation of view-based access control model identification and cash device, control system and method

Technical field

The valuation identified the present invention relates to a kind of automatic cash register system more particularly to a kind of view-based access control model and cash device, Control system and method.

Background technique

Chinese about four trillion yuan of food and drink market scale in 2018, wherein group's meal accounts for 1/3, about 1,300,000,000,000 yuan, the market space is huge Greatly.But it is risen (cashier, the person of ordering) by cost of labor, Customer Experience bad (clearing that stand in a long queue, clearing are error-prone), consumption number According to the restriction for retaining less and converting low (being in great demand and unsalable, meal time, consuming frequency, visitor's unit price) three big main problems, entirety For, the profitability of group meal industry businessman is poor.Businessman is badly in need of effective scheme, to help to improve the level of profitability and operation effect Rate.

Customer is free, self-help settlement, and self-help payment is the good scheme to solve the above problems.

Customer is free, is a kind of change of cafeteria mode, is mainly characterized by service staff and (is specifically responsible for beating The waiter of dish, hereinafter referred to as beats dish person) dish is dispensed into pickles dish according to scheduled component in advance according to target price, it cares for When visitor has dinner, according to oneself hobby and budget, the corresponding vegetable of self-service selection is into pallet；Cashier fixes a price according to vegetable and counts The total price of all vegetables on pallet is calculated, and inputs total price on cash register, customer can be by cash payment, or passes through one Cartoon payment completes payment perhaps by barcode scanning payment or other e-payment means.

In above-mentioned process, it will be seen that reducing customer by free and beating the time for communication of dish person, customer Selection with beat dish person benefit dish can it is synchronous carry out, improve efficiency；And in settlement process, cashier needs to identify and remember The price for firmly corresponding to vegetable completes the input of price and the calculating of total price, needs to expend time also easy error；And in payment ring Section, cash are related to small change, inefficiency and unhygienic, are gradually eliminated；The mode of e-payment, the interaction for the medium being related to, Such as dig out card or dig out mobile phone and recall two dimensional code, it requires to expend the time, it is inefficient.

With the development of information technology and technology of Internet of things, it also proposed some solution party for settling problems on the market Case is mainly segmented into two major classes:

The first kind: tableware identification

This kind of scheme is identified and is fixed a price by tableware feature when main, to allow cash register that can identify correspondence Vegetable and identify corresponding price, to complete to settle accounts.

Specific means in advance are segmented into two kinds again:

By electronic tag, such as rfid chip or nfc chip are implanted into tableware, and price is written in the chips, Cash register reads the corresponding price of all chips on pallet in clearing, to calculate total by chip information read-write equipment Valence；The major drawbacks of such scheme are to rely on chip, since increase cost, two carry out chip in tableware, by long-term point It picks, sterilizes, can be reduced by collision and high temperature, the reliability of chip, to occur misidentifying or leaking identification；And these feelings Condition is not easy to be found, and leads to the mistake of clearing, and can not be traced.

Pass through vision prescription, using special shape, color, or the tableware of label；It is corresponding that each tableware is provided in advance Price, so that the vegetable that the tableware is contained sells corresponding price；This kind of scheme, in previous implementation, dependent on meal The preparatory training classification of tool needs re -training and classification when dining room needs to be added new tableware or replacement tableware, Several days time is expended, flexibility is poor, influences the normal operation in dining room.

Second class: vegetable identification

This kind of scheme mainly utilizes the feature such as shape of food itself, and color is identified and settled accounts；Such as CN104077842 carries out off-line training and study to vegetable using the method for machine learning, is then identified in clearing Obtain corresponding dish valence；Such methods main problem is that using off-line training method, need vegetables all in the dining room It collects in advance and is trained classification；And the same vegetable, the way of every day may be different, moreover, dining room is sought after New dish is released at any time；It is not possible that having time and manpower are for special acquisition data and training in advance；So the program is not one Effective scheme.

With the development of information technology and technology of Internet of things, it also proposed some solution party for payment problem on the market Case is mainly segmented into three categories:

The first kind: cCredit card payment

The program depends on electronic card, is easy to appear situations such as leaving behind card or card wreck；And it was paying Cheng Zhong needs to find out card and swipes the card, needs to expend the time.

Second class: barcode scanning payment

Barcode scanning payment be it is universal with smart phone and mobile Internet, be proposed a kind of new payment scheme；Mainly Advantage is for identification code to be integrated in mobile phone app, can complete to pay as long as with mobile phone, avoids and leaves behind card and card damage Bad problem, the program main problem is that, payment needs first to dig out mobile phone every time, then recalls app and recalls two dimensional code, needs Expend the time longer than cCredit card payment.

Third class: brush face payment

With the development of face recognition technology, some enterprises are proposed brush face payment function, and customer is by facial picture and account Family binding exports the account of oneself and brush face is verified, then can complete to pay in payment；The advantage of the program is matchmaker It is situated between；Efficiency is improved to a certain extent；But each account number that still input oneself or short password, it is still not square enough Just；In technology realization, the program is extensive there is no realizing, the 1:N face alignment of high accuracy, but by account, it is real The people for having showed 1:1 compare, and in this implementation, face is only as the cryptographic check of verification, rather than identification；And by short The mode of password, and local face database is navigated within the scope of the personnel of very little by the screening of short password, the order of magnitude Tens；The realization of the technology is relatively very simple and not no technical progress.

Generally, for the scene in current free dining room, still lack a kind of clearing and efficient enough, the embodiment party of payment Just and it is reliable and stable, while the scheme that econmics comparison is good.

Summary of the invention

For overcome the deficiencies in the prior art, the purpose of the present invention is to provide the valuation and receipts of a kind of identification of view-based access control model The silver device, control system and method.

The purpose of the present invention is implemented with the following technical solutions:

A kind of valuation of view-based access control model identification and cash device, including host, customer's touch screen, vegetable identification camera and Service plate placement plate, the vegetable identify that service plate placement plate described in camera face is set to the lower section of customer's touch screen, the meal Disk placement plate is fixed on cashier, and customer is arranged above the display area of customer's touch screen and identifies camera, In the lower section of customer's touch screen, setting vegetable identifies that secondary light source, the vegetable identification secondary light source are arranged in the vegetable Identify the two sides of camera.

Preferably, device further includes auxiliary multifunctional brush card machine, more by the auxiliary after recognition of face, which is paid, to fail The function POS completes payment.

Preferably, in the bottom surface of service plate placement plate setting pressure sensor to perceive the vegetable for whether needing to be settled accounts pendulum It is placed in service plate placement plate.

Preferably, the vegetable identification camera changes according to the image of input, and whether automatic sensing needs the dish settled accounts Product are placed in service plate placement plate.

Preferably, the service plate placement plate is provided with luminous zone, when the vegetable wait settle accounts is placed in service plate placement plate, Luminous zone is blocked, to enhance the standard that the vegetable whether vegetable identification camera perception needs to be settled accounts is placed in service plate placement plate True rate.

Preferably, device further includes customer service touch screen, and the customer service touch screen and the opposite setting of customer's touch screen are in institute State the two sides of host.

Preferably, device further includes voice announcer, from trend customer broadcast voice prompting, voice prompting content include to Whether payment vegetable detail, dish valence, clearing total amount and clearing succeed.

Preferably, the customer service touch screen and customer's touch screen use liquid crystal touch screen, and the vegetable identification secondary light source is adopted Use LED light source.

A kind of valuation and cash register control system of view-based access control model identification, control system includes: vegetable recording module, Gu Kexin Cease recording module and clearing payment module.

The vegetable recording module identifies that camera takes pictures to the new vegetable in standard service plate by vegetable, will acquire vegetable Photo sends processor to, extracts image feature information to vegetable photo by the processor of host and combines artificial treatment, determines Menu name and price；

The Customer Information recording module identifies that camera takes pictures to customer head by customer, the customer that will acquire Photo sends processor to, carries out graphic feature extraction to customer's photo by the processor of host and is shown in customer's touch screen On, and combine customer's artificial treatment, typing customer face, identity information and payment accounts information；

The clearing payment module takes pictures to the vegetable in service plate placement plate by vegetable identification camera, is valuated simultaneously Dining information is shown on customer's touch screen；Meanwhile identifying that camera captures customer face photo by customer, through processor point Analysis identification customer simultaneously shows Customer Information on customer's touch screen；Show that customer confirms branch simultaneously on customer's touch screen The option paid is confirmed by customer completes the payment of brush face.

Preferably, system further include auxiliary payment module, the auxiliary payment module by assist multifunctional brush card machine into Row gathering, the auxiliary payment module includes IC cCredit card payment mode, two dimensional code payment mode or quick flashing payment mode.

A kind of valuation and cash method of view-based access control model identification, method include:

Preparation process, preparation process complete the typing of tableware template, are that vegetable typing and the vegetable identification in use process are done It is good to prepare；And Customer Information typing is completed, the face information of customer and account information are bound, carry out standard for the payment of brush face It is standby；

Meal use process is opened, opening meal use process includes clearing payment when opening vegetable typing before the meal, opening meal, Yi Jikai Postprandial data check process；Wherein vegetable typing will update and adjustment history vegetable and price, while the new vegetable of typing and right Dutiable value；Clearing payment realizes clearing in such a way that the vegetable of vision and vegetable identification combine, and by exempting from close 1: The mode of the brush face payment of N is paid；Data are checked, the sales situation of every meal is counted, sales volume, sale including each dish Total value, customer's personal consumption volume, customer's history consumption vegetable and customer's history consume total value；

Vegetable and nutrition data push process, by customer for a period of time in dining information, dining ingredient and nutrition data, Consumption information and healthy pantry advisory information send customer to.

Preferably, the tableware template typing of the preparation process includes the following steps:

(1) power on, opening device；So that vegetable identification secondary light source beats bright service plate placement plate, cooperation camera ginseng Number obtains clear photograph；There is flexibly stable adaptability photo to obtain environment in different dining rooms with this；

(2) operator gets out tableware all in dining room, and prepares a certain amount of vegetable；

(3) operator's activation system carries out tableware shooting；

(4) operator takes a tableware, and a certain amount of vegetable is packed into tableware, and tableware is placed on each of pallet Position, and the angle of tableware is adjusted, shoot several photos；Photo is stored in same file position；

(5) (4) operator repeats step to each tableware；

(6) after all tableware pictures are ready to complete, run the typing of tableware template and learning program, complete tableware typing and Study；

(7) the tableware of typing can be used in the normal operation in subsequent dining room；When replacement or increase new meal When tool, then delete should not tableware picture, the picture of the new tableware of retake simultaneously repeats (6) typing and study that step completes new tableware.

Preferably, the Customer Information typing of the preparation process includes the following steps:

(1) opening device, activation system, selection enter Customer Information recording module；

(2) customer according to system suggestion inputs account information, and account information can be with the cell-phone number of association user, body on backstage Part card number, student number, mailbox, address, and/or work number；

(3) after customer inputs information, whether system check is legal account, and confirmation account whether there is；If not conforming to rule Prompt requires to re-enter；

(4) prompt customer faces customer and identifies camera, and prompting mode includes voice prompting, text prompt, and/or instruction Lamp prompt；

(5) starting camera program captures client head photo, and prompt customer's new line is nodded, left and right shakes the head, and/or blinks, After the completion of photograph taking, it is sent into photographic quality program, photographic quality is assessed by assessment algorithm, first determines whether that photo is It is no fuzzy, it takes pictures again if fuzzy, human face characteristic point is extracted by Feature Extraction algorithm if photographic quality passes through, is sentenced Whether disconnected face is complete, whether angle is suitable, undesirable, takes pictures, meets the requirements again, carries out in next step；

(6) remind customer to identify successfully, and the picture of angles multiple in shooting process is stored in the association picture position of the customer It sets；Extract the face characteristic of all angles by face characteristic extraction algorithm to the photo of customer, and by these features and the Gu The account information association of visitor；

(7) typing customer account password, pop-up input password window simultaneously prompt input password, the following payment of password reply When special circumstances；Prompt inputs password again and verifies；Not by then re-entering, by then carrying out in next step；

(8), if face is crossed in typing before in the account, judge whether new typing face and original face are same One people；If it is pass through；Otherwise report an error false information, and requires connection staff；

(9) it is fully completed, prompts typing success.

Preferably, the vegetable typing for opening meal use process includes the following steps:

(1) opening device, activation system, selection enter vegetable recording module；It opens vegetable identification secondary light source and plays bright service plate Placement plate cooperates camera parameter, obtains clear photograph；There is flexibly stable adaptability photo to obtain in different dining rooms with this Take environment；

(2) operator gets out a sample of each vegetable of the meal, and each vegetable is protected in a tableware This vegetable is demonstrate,proved to remain loaded into the tableware；

(3) the vegetable for needing typing is placed into standard service plate by operator, and standard service plate is placed into service plate placement plate Cog region；

(4) after confirming correct placement, vegetable recognizer is started；Recognizer calls vegetable identification camera to shoot one Photo is sent into vegetable detection program after verifying quality；The vegetable and sundries in Programmable detection picture are detected, and by vegetable in picture It is split from picture with sundries part, is sent into sort program, sundries is rejected；

(5), for the vegetable detected, tableware used in the vegetable is identified first；Tableware information is extracted, with typing before Tableware template compares, and matches highest several tablewares；If matched highest tableware meets threshold condition, calculate and The similarity of second similar tableware；Otherwise it is assumed that this tableware is not present in the tableware template for realizing typing, this meal Tool needs typing tableware template；If tableware exists in a template, and most like and time similar difference is bigger, then determining should Tableware is the similar tableware of top1 in tableware template；Otherwise, it dishes out several possible tablewares, allows operator's artificial selection, with true Definitely exactness；The above matching process for completing tableware；

(6) the matched vegetable of tableware is completed for confirmation, a text box that can be inputted occur；Operator passes through touch screen Text frame is clicked, the title and price of the vegetable can be inputted；

(7), for the vegetable of each completion price setting, picture is sent into vegetable recognizer, passes through the vegetable of precondition Feature Extraction Algorithm extracts the characteristic of the vegetable；

(8) operator clicks confirmation vegetable typing, and whether all vegetables of program ver-ify have price, and identical dish The not different price of product, then pass through；All vegetables are entered into the vegetable database of the meal on the standard service plate；

(9) repeat the above process (3)~(8), complete the typing of all vegetables.

Preferably, the clearing payment for opening meal use process includes the following steps:

(1) opening device, activation system, selection enter clearing payment module；It opens vegetable identification secondary light source and plays bright service plate The corresponding cog region of placement plate has flexibly stable adaptability photo to obtain environment in different dining rooms with this；

(2) the standard service plate for filling vegetable tableware is placed into service plate placement plate by client, and vegetable identifies that camera shoots standard The photo of service plate, and send photo to detection program；

(3) the vegetable and sundries in Programmable detection photo picture are detected, by Recognize algorithm trained in advance by picture Middle vegetable and sundries part are divided from picture, are sent into sort program, sundries is rejected；

(4), for the vegetable detected, the vegetable frame choosing that will test out is divided；

(5) tableware feature and the meal are recorded some vegetable, first extraction tableware feature by the vegetable identification range confirmed The tableware Characteristic Contrast entered obtains several tablewares before similarity；If top1 tableware is unsatisfactory for similarity by threshold value, sentence The fixed tableware is illegal；If top1 crosses threshold value, the value of top1-top2 is calculated；If the value of top1-top2 is greater than threshold value, Then the tableware of top1 has conspicuousness, then only need to identify vegetable in the corresponding tableware of top1；If the value of top1-top2 is less than threshold Value, then the tableware of top1 does not have conspicuousness, corresponding vegetable need to be matched in the corresponding tableware of top1 and top2, to ensure standard Determining and reliability；

(6) after the vegetable identification range of confirmation, vegetable is identified: extraction vegetable feature first, vegetable feature and corresponding tableware Vegetable Characteristic Contrast in range, obtains vegetable sequencing of similarity；If top1 vegetable is unsatisfactory for similarity by threshold value, sentence The fixed non-typing of the vegetable；If top1 crosses threshold value, the value of top1-top2 is calculated；If the value of top1-top2 is greater than threshold value, Then the vegetable of top1 has conspicuousness, then is determined as the vegetable of top1；If the value of top1-top2 is less than threshold value, the dish of top1 Product do not have conspicuousness, the match information of comprehensive dish and the match information of tableware, and weighted calculation obtains an integrated value, and takes Top1 vegetable is the vegetable identified；Risk is recorded simultaneously, in case retrospect；

Repeat step (5)~(6), complete the identification of all vegetables, and according to the price of typing, obtain the price of vegetable, Calculate dining total price；

(8) above-mentioned photo with partitioning scheme and shows on customer's touch screen dutiable value, and show dining total price, text Word and/or voice prompting customer dining total price are simultaneously paid；

After vegetable completes valuation or valuation simultaneously, starting customer identifies camera, obtains and cleans complete customer's head Portion's photo extracts human face characteristic point, judges whether face is complete, and whether angle is suitable, improper, retakes, properly then carries out down One step；

(10) the face picture for meeting quality is sent into recognition of face program, extracts face characteristic, and and logging data Face characteristic in library is compared；Obtain the human face photo of similarity top1 and top2；First determining whether the similarity of top1 is It is no to meet by threshold value, if conditions are not met, then determining the customer not in user library；Prompt the non-typing of face；Typing people at this time Face is paid by the mode of auxiliary payment；

If the face of top1 meets threshold value, needing judgement is the risk of this people；The value of top1-top2 is calculated, such as The fruit value is greater than risk threshold value, then determines that top1 has conspicuousness；If the value is less than risk threshold value, determine that top1 has wind It is dangerous, need additional verifying means to guarantee accurate fixed, at this point, typing password when customer can input typing face carries out two Secondary verification completes identity validation, and carries out brush face payment.

Preferably, the postprandial data for opening meal use process check to include: the total pin in dining room counted in a period of time Sell the sales data of data and single vegetable, dining consumption total price, dining ingredient and the nutrition data of each customer.

Compared with prior art, the beneficial effects of the present invention are: using more basis clearing structure, system and method So that clearing and payment are efficiently, but also the overall structure of device is simplified；Vegetable Input Process is simple and efficient；User face Feature and user account are directly linked, without in addition inputting the account of user or in addition providing user account institute when brush face is paid The ancillary equipments such as mobile phone, the card of binding facilitate the user of both hands end service plate to settle accounts；The recognition accuracy of whole system is high, real It is convenient and reliable and stable to apply, and econmics comparison is good.

Detailed description of the invention

Fig. 1 is the valuation of view-based access control model identification and the structural schematic diagram of cash device；

Fig. 2 is the valuation of view-based access control model identification and the main interface schematic diagram of cash register system；

Fig. 3 is the interface schematic diagram of Customer Information recording module；

Fig. 4 is the interface schematic diagram of vegetable recording module；

Fig. 5 is the interface schematic diagram for settling accounts payment module；

Fig. 6 is Customer Information input method flow chart；

Fig. 7 is vegetable input method flow chart；

Fig. 8 is the method flow diagram of clearing payment.

In figure: 100, cashier；1, host；2, customer service touch screen；3, customer's touch screen；4, customer identifies camera；5, Vegetable identifies camera；6, service plate placement plate；7, multifunctional brush card machine is assisted；71, input keyboard；72, credit card area；8, standard meal Disk；9, vegetable identifies secondary light source.

Specific embodiment

With reference to the attached drawing in the embodiment of the present invention, technical solution in the embodiment of the present invention carries out clear, complete Ground description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Based on this Embodiment in invention, every other reality obtained by those of ordinary skill in the art without making creative efforts Example is applied, shall fall within the protection scope of the present invention.

It should be noted that it can be directly on another component when component is referred to as " being fixed on " another component Or there may also be components placed in the middle.When a component is considered as " connection " another component, it, which can be, is directly connected to To another component or it may be simultaneously present component placed in the middle.When a component is considered as " being set to " another component, it It can be and be set up directly on another component or may be simultaneously present component placed in the middle.Term as used herein is " vertical ", " horizontal ", "left", "right" and similar statement for illustrative purposes only.

Unless otherwise defined, all technical and scientific terms used herein and belong to technical field of the invention The normally understood meaning of technical staff is identical.Term as used herein in the specification of the present invention is intended merely to description tool The purpose of the embodiment of body, it is not intended that in the limitation present invention.Term " and or " used herein includes one or more phases Any and all combinations of the listed item of pass.

Referring to Fig. 1, a kind of valuation of view-based access control model identification and cash device, including host 1, customer service touch screen 2, customer's touching Control screen 3, vegetable identification camera 5, service plate placement plate 6 and auxiliary multifunctional brush card machine 7, the customer service touch screen 2 and customer's touching Control screen 3 is opposite to be staggered on cashier 100.Service plate placement plate 6 described in vegetable identification 5 face of camera is set to Gu The lower section of objective touch screen 3.Service plate placement plate 6 is fixed on cashier 100, above the display area of customer's touch screen 3 Customer is set and identifies camera 4, setting vegetable identifies secondary light source 9, the vegetable identification in the lower section of customer's touch screen 3 The two sides of the vegetable identification camera 5 are arranged in secondary light source 9.

Wherein, customer service touch screen 2 and customer's touch screen 3 use liquid crystal touch screen, and the vegetable identification secondary light source 9 uses LED light source.

Wherein, the vegetable identification camera 5 changes according to the image of input, and whether automatic sensing needs the vegetable settled accounts It is placed in service plate placement plate 6.

Selectable, service plate placement plate 6 is provided with luminous zone, when the vegetable wait settle accounts is placed in service plate placement plate 6, Luminous zone is blocked, is placed in service plate placement plate 6 with enhancing the vegetable whether the vegetable identification perception of camera 5 needs to be settled accounts Accuracy rate.

For assisting multifunctional brush card machine 7, including input keyboard 71, credit card area 72, two-dimensional code scanning area and/or quick flashing branch Pay induction zone.It two-dimensional code scanning area can be with means of payment such as counterpart expenditure treasured or wechats.Thus research and development cost is reduced, is not happened suddenly Cannot brush face payout status do emergency payment processing.

Further, device further includes voice announcer, broadcasts voice prompting from trend customer, voice prompting content includes Whether vegetable detail, dish valence, clearing total amount and clearing to be paid succeed.

In addition, an also settable dedicated terminating machine in dining room, is recorded with the information that this is realized with cashier mutually separates Enter.

- Fig. 5 referring to fig. 2, invention additionally provide the valuation and cash register control system of a kind of view-based access control model identification, control system It include: vegetable recording module, Customer Information recording module and clearing payment module.

Wherein, referring to Fig. 3, Customer Information recording module identifies that camera 4 takes pictures to customer head by customer, will Customer's photo of acquisition sends processor to, is carried out graphic feature extraction to customer's photo by the processor of host 1 and is shown in On customer's touch screen 3, and combine customer's artificial treatment, typing customer face, identity information and payment accounts information.

Wherein, referring to fig. 4, vegetable recording module identifies that camera 5 claps the new vegetable in standard service plate 8 by vegetable According to will acquire vegetable photo and send processor to, extracted and image feature information and combined to vegetable photo by the processor of host 1 Artificial treatment determines menu name and price.

Wherein, referring to Fig. 5, payment module is settled accounts by vegetable and identifies that camera 5 carries out the vegetable in service plate placement plate 6 It takes pictures, valuate and show dining information on customer's touch screen 3；Meanwhile identifying that camera 4 is captured customer face and shone by customer Piece analyzes identification customer through processor and shows Customer Information on customer's touch screen 3；It is same on customer's touch screen 3 When show customer confirm payment option by customer confirm complete brush face payment.

In addition, system further includes auxiliary payment module, auxiliary payment module is by assisting multifunctional brush card machine 7 to be received Money, auxiliary payment module includes IC cCredit card payment mode, two dimensional code payment mode or quick flashing payment mode etc..

A kind of valuation and cash method of view-based access control model identification, method include:

Preparation process, preparation process complete the typing of tableware template, are that vegetable typing and the vegetable identification in use process are done It is good to prepare；And Customer Information typing is completed, the face information of customer and account information are bound, carry out standard for the payment of brush face It is standby.

Meal use process is opened, opening meal use process includes clearing payment when opening vegetable typing before the meal, opening meal, Yi Jikai Postprandial data check process；Wherein vegetable typing will update and adjustment history vegetable and price, while the new vegetable of typing and right Dutiable value；Clearing payment realizes clearing in such a way that the vegetable of vision and vegetable identification combine, and by exempting from close 1: The mode of the brush face payment of N is paid；Data are checked, the sales situation of every meal is counted, sales volume, sale including each dish Total value, customer's personal consumption volume, customer's history consumption vegetable and customer's history consume total value.

Vegetable and nutrition data push process, by customer for a period of time in dining information, dining ingredient and nutrition data, Consumption information and healthy pantry advisory information send customer to.

Wherein, tableware template Input Process are as follows:

Power on, opening device；Computer in equipment, camera, the equipment such as light source start simultaneously；Light compensating lamp is main For beating bright pallet area, cooperates camera parameter, obtain clear photograph；Can have in this way to different dining room environment flexibly steady Fixed adaptability.

Operator gets out tableware all in dining room, and prepares a certain amount of vegetable, and vegetable can be random.

Operator starts tableware photographing program.

Operator takes a tableware, and a certain amount of vegetable is packed into tableware, and tableware is placed on to each position of pallet It sets, and adjusts the angle of tableware, shoot several photos；And these photos are stored in same file position.

Operator repeats aforesaid operations to each tableware.

After all tableware pictures are ready to complete, the typing of tableware template and learning program are run, in several seconds kinds, tableware will be complete At typing and study.

The tableware of typing can be used in the normal operation in subsequent dining room；If replacement increases new meal Tool, then need to only delete should not tableware picture or retake take the photograph the picture of new tableware, and operation tableware recording program is again It can.

Above step relies on the acquisition and algorithm training of prior high-volume tableware data, so that obtaining one has pumping The intelligent algorithm model of pick-up tool feature；

Wherein, Customer Information typing, flow diagram shown in Figure 6: the purpose of face typing is by the people of customer Face information and account information binding, comprising:

Start Customer Information recording program.

Customer according to prompt input account information, account information backstage can with the cell-phone number of association user, identification card number, Work number etc.；So user can input associated any information.

After user inputs information, whether program ver-ify is legal account first, i.e. account whether there is；If not conforming to rule It is required that re-entering.

The formal camera of user is prompted, prompting mode can pass through voice, text, the diversified forms such as indicator light and combination.

Starting camera program, which is grabbed, to be taken a picture, and when taking pictures, light compensating lamp can be automatic according to the light and shade of ambient conditions and face Brightness is adjusted, to take clearly human face photo；After the completion of photograph taking, it is sent into photographic quality assessment algorithm module, first Judge whether photo obscures, clapped again if fuzzy, by if by Feature Extraction algorithm, extract human face characteristic point, Judge whether face is complete, whether angle is suitable；It is improper, it retakes；It is suitable then carry out in next step.

Customer or so is reminded to shake the head；It takes pictures and judges whether quality and angle meet；Calculating angle has a set of advanced herein Algorithm.

It reminds customer to come back to nod；It takes pictures and judges whether quality and angle meet；Calculating angle has a set of advanced herein Algorithm.

Above-mentioned steps all pass through, then customer is reminded to identify successfully, and should by the deposit of the picture of angles multiple in shooting process The association Pictures location of customer；And start the algorithm for extracting face characteristic, the face characteristic of all angles is extracted, and these are special Sign is associated with the account of the customer.

Pop-up input password window simultaneously prompts input password.

User can input a string of passwords, to cope with special circumstances when the following payment；Under normal circumstances, user is paying When do not need to use this password.

Prompt inputs password again and verifies；Not by then re-entering, by then carrying out in next step.

If face is crossed in typing before in the account, need to judge whether are new typing and face and original face It is the same person；If it is pass through；Otherwise report an error false information, and requires connection staff.

It is fully completed, prompts typing success.

In order to cope with the similar situation of multiple people that may be present；The present invention provides a kind of methods of dynamic threshold, i.e., Spy is carried out for the face information of each new typing, first extraction feature, and by the other users in this feature and face database Sign comparison, seeks similarity；If there are the higher situations of similarity for the face and other faces, to this face setting one It is a relatively high to pass through threshold value；It is i.e. following into others is known, it needs to be determined as with the human face similarity degree very Gao Caineng same People；, whereas if the similarity of the face and other faces is not high, then one can be set to the face and relatively low is passed through Threshold value；The setting method of threshold value specific value may be implemented there are many algorithm；The method of dynamic threshold is mainly taken into account and safe Property and experience.

On realizing route, the step of face typing, it can be realized in cash register robot the machine, it can also be in dining room The dedicated terminating machine of setting one is realized, can also be realized by mobile terminals such as mobile phones.

Wherein, vegetable typing flow diagram shown in Figure 7, includes the following steps.

Power on, opening device；Computer in equipment, camera, the equipment such as light source start simultaneously；Light compensating lamp is main For beating bright pallet area, cooperates camera parameter, obtain clear photograph；Can have in this way to different dining room environment flexibly steady Fixed adaptability.

Operator gets out a sample of each vegetable of the meal, and each vegetable guarantees in a tableware This vegetable remains loaded into the tableware in the meal.

The present invention supports the disposable multiple vegetables of typing to improve efficiency.

Multiple vegetables for needing typing are placed into pallet by operator, and pallet is placed into the knowledge on tempered glass table top Other area.

After confirming correct placement, start recognizer；Recognizer calls camera to shoot a photo, after verifying quality It is sent into vegetable and detects program；Detect the vegetable and sundries in Programmable detection picture, and by vegetable in picture and sundries part from figure It is split in piece, is sent into sort program, sundries is rejected and (relies on the algorithm that can identify sundries of a training in advance herein Model).

For the vegetable detected, tableware used in the vegetable is identified first；Extract tableware information, the meal with typing before Tool template compares, and matches highest several tablewares；If matched highest tableware meets certain threshold condition, count Calculate the similarity with second similar tableware；Otherwise it is assumed that this tableware is not present in the tableware template for realizing typing, this A tableware needs typing tableware template；If tableware exists in a template, and most like and time similar difference is bigger, then sentences The fixed tableware is the similar tableware of top1 in tableware template；Otherwise, it dishes out several possible tablewares, allows operator's artificial selection, To ensure accuracy；The above matching process for completing tableware.

The matched vegetable of tableware is completed for confirmation, a text box that can be inputted occurs；Operator passes through touch screen point Text frame is hit, the price of the vegetable can be inputted.

For the vegetable of each completion price setting, picture is sent into vegetable recognizer, extracts the characteristic of the vegetable； The algorithm model that can extract vegetable feature of a precondition is relied on herein；The model is by acquiring a large amount of vegetable number According to, after classification by deep learning training obtain.

Operator clicks confirmation typing, and whether all vegetables of program ver-ify have price, and identical vegetable does not have Different prices, then pass through；All vegetables are entered into the vegetable database of the meal on the service plate.

It repeats the above process, completes the typing of all vegetables of the meal.

Wherein, clearing method of payment mainly includes the following steps referring to flow diagram described in Fig. 8.

Power on, opening device；Computer in equipment, camera, the equipment such as light source start simultaneously；Light compensating lamp is main For beating bright pallet area, cooperates camera parameter, obtain clear photograph；Can have in this way to different dining room environment flexibly steady Fixed adaptability.

Starting clearing program.

Since clearing program will realize noninductive automatic identification, the interaction of customer is reduced as far as possible；So, it is necessary first to intelligence It is clear to one that energy ground is captured, completely, and unobstructed service plate picture.

Vegetable identifies camera starting, and program intelligent candid one is opened service plate picture and identified.And it is sent into the knowledge in program Other area judgment module.

In order to guarantee the complete of service plate picture, cog region devises a sizeable identification frame, to ensure the identification For frame in the case where being blocked completely, in obtained photo, service plate is complete；Cog region can also by pressure sensor, The methods of infrared distance sensor is prior；Program judges that cog region is completely covered at this time, is determined with service plate and is correctly placed in knowledge In other area；If service plate covering part cog region, program will determine that service plate is not correctly placed on cog region.

It is clear and unobstructed in order to ensure service plate picture, therefore, to assure that the service plate in photo is stationary, and is eaten The object also not moved above disk, such as hand；In specific implementation, program can be implemented every taking a picture for a period of time On, it can also all be taken with every frame, in order to realize speed, can also preferably select to capture every a few tens of milliseconds primary；It is capturing to figure It is compared after piece with picture that front is captured, if picture is without difference, then it is assumed that service plate is stable and does not block； If picture is different with respect to the picture of front, determine that service plate is stable not yet or service plate above have mobile block Object.

The vegetable and sundries in Programmable detection picture are detected, and vegetable in picture and sundries part are partitioned into from picture Come, be sent into sort program, sundries is rejected and (relies on the algorithm model that can identify sundries of a training in advance herein).

For the vegetable detected, the vegetable frame choosing that will test out is divided.

For some vegetable, extraction tableware feature first obtains the tableware Characteristic Contrast of tableware feature and meal typing Several tableware before similarity；If top1 tableware is unsatisfactory for similarity by threshold value, determine that the tableware is illegal；If Top1 crosses threshold value, then calculates the value of top1-top2；If the value of top1-top2 is greater than threshold value, it is aobvious to illustrate that the tableware of top1 has Work property then only needs to identify dish in the corresponding tableware of top1；If the value of top1-top2 is less than threshold value, illustrate The tableware of top1 does not have conspicuousness, needs to match corresponding dish in the corresponding tableware of top1 and top2, accurately fixed to ensure And reliability.

After the vegetable identification range of confirmation, vegetable is known similar otherwise with tableware knowledge otherwise；Dish is extracted first Product feature, vegetable feature and the vegetable Characteristic Contrast within the scope of corresponding tableware obtain several vegetables before similarity；If top1 Vegetable is unsatisfactory for similarity by threshold value, then determines the non-typing of the vegetable；If top1 crosses threshold value, calculate top1-top2's Value；If the value of top1-top2 is greater than threshold value, illustrates that the vegetable of top1 has conspicuousness, be then determined as the vegetable of top1；If The value of top1-top2 is less than threshold value, then illustrates that the vegetable of top1 does not have conspicuousness, at this time, it may be necessary to the match information of comprehensive dish With the match information of tableware, weighted calculation obtains an integrated value, and taking top1 dish is the dish identified；Wind is recorded simultaneously Danger, in case retrospect；By method above-mentioned, the accuracy that vegetable identification may be implemented reaches 99.99% or more.

Preceding step is repeated, completes the identification of all vegetables, and according to the price of typing, obtain the price of vegetable, is calculated Total price out.

Specifically in realization, tableware identification, vegetable identification can be carried out individually, can also be in conjunction with progress；If only used Tableware then fixes a price to tableware, and the vegetable to dutiable value is packed into operation；Another realizing route can also not depend on Tableware directly carries out the identification of vegetable.

By text importing total price, price is broadcasted by prompt tone, customers aware and is confirmed with allowing；If customer has found There is identification or calculate mistake, can be helped by the searching on touch screen, calling cashier carries out manual intervention processing, to improve Reliability.

After vegetable completes valuation, starting recognition of face camera, camera program, which is grabbed, to be taken a picture, when taking pictures, light filling Lamp can be according to the light and shade automatic regulating lightness of ambient conditions and face, to take clearly human face photo；Photograph taking is completed Afterwards, be sent into photographic quality assessment algorithm module, first determine whether photo obscures, clapped again if fuzzy, by if pass through Feature Extraction algorithm extracts human face characteristic point, judges whether face is complete, whether angle is suitable；It is improper, it retakes；It closes It is suitable then carry out in next step.

The face picture that quality will be met is sent into recognition of face program, extracts face characteristic, and and input database In face characteristic be compared；Obtain the face of most like top1 and top2；First determine whether the similarity of top1 is full Foot is by threshold value, and particularly, threshold value herein is the preferred dynamic threshold come out according to calculated in advance in face database Value；If conditions are not met, then determining the customer not in user library；Prompt the non-typing of face；With typing face or it can lead at this time Other modes are crossed to pay.

If the face of top1 meets threshold value, needing judgement is the risk of this people；The value of top1-top2 is calculated, such as The fruit value is greater than risk threshold value, and (special, risk threshold value herein is the face database according to typing, calculated Preferred dynamic threshold), then determine that top1 has conspicuousness；If the value is less than risk threshold value, determine that top1 has risk Property, need additional verifying means to guarantee accurate fixed, at this point, the password progress of typing is secondary when customer can input typing face Identity validation is completed in verification.

By three above step, for most of people, the accuracy of recognition of face may be implemented to 99.999% More than, in addition cryptographic check, then may be implemented the accurate of recognition of face 100%；Realize that user's is quick in most cases Reliable 1:N's exempts from close brush face payment.

In addition, it includes: the total sales data in dining room counted in a period of time that data, which check that the i.e. postprandial data of process are checked, With the sales data of single vegetable, dining consumption total price, dining ingredient and the nutrition data of each customer.

In meal or postprandial, dining room manager can check statistical data；The pin of every meal can be checked in the machine Data are sold, web page can also logged in from the background, check the sales data of the every meal of history.Tune is made to vegetable ingredient etc. with this It is whole.

In addition, system and method additionally provides Premium Features.

For vegetable, a vegetable inventory can be arranged according to the traffic-operating period in dining room in advance；Simultaneously for each dish Product, can typing vegetable in advance ingredient data and nutrition data.

Under the premise of in front, when every meal opens meal logging data, corresponding vegetable can be chosen, in this way, then settling accounts When, it can show the nutrition data of menu name and vegetable.

Under basis in front, the quantity of each raw material consumed daily can be counted；Do not have dining room predetermined raw material in advance Data reference is provided.

On the basis of in front, everyone every meal, the nutrition intake situation daily or in a period of time can be counted； And corresponding Health & Fitness Tip is provided with this；The data and suggestion can be sent in the account of customer by mobile use；Customer Mobile terminal and mobile process, real time inspection can be passed through；And the push for the prompting that can secure good health.

Technical solution of the present invention is with the artificial main body of intelligent cashier machine, in conjunction with dining room layout, standard service plate, data The total solution of platform.By the identification of automatic vegetable, exempt from the functions such as close brush face payment, digital intelligent operation, solves The practical problems such as the custom queueing time is long in actual scene, settlement efficiency is low, vegetable data difficulty is retained improve client's body to realize It tests, reduces queuing time, precisely analyzes data, a variety of client's actual demands such as cut operating costs.

The invention adapts to but the dining place such as be not limited to dining room, school lunch service.

The aforementioned description to specific exemplary embodiment of the invention is in order to illustrate and illustration purpose.These descriptions It is not wishing to limit the invention to disclosed precise forms, and it will be apparent that can much be changed according to appeal introduction And variation.The purpose of selecting and describing the exemplary embodiment is that explaining specific principle of the invention and its actually answering With so that those skilled in the art can be realized and utilize a variety of different exemplary implementation schemes of the invention and Various chooses and changes, i.e., according to the above description of the technical scheme and ideas, those skilled in the art can be done Various other corresponding changes and deformation out, and all these change and deformation all should belong to the claims in the present invention Protection scope within.

Claims (16)

1. valuation and the cash device of a kind of view-based access control model identification, including the identification camera shooting of host (1), customer's touch screen (3), vegetable Head (5) and service plate placement plate (6), the vegetable identify that the customer that is set to of service plate placement plate (6) described in camera (5) face touches The lower section of control screen (3), it is characterised in that: the service plate placement plate (6) is fixed on cashier (100), in the customer Setting customer identifies camera (4) above the display area of touch screen (3), and dish is arranged in the lower section of customer's touch screen (3) Product identify secondary light source (9), and vegetable identification secondary light source (9) setting is in the two sides of vegetable identification camera (5).

2. the apparatus according to claim 1, it is characterised in that: device further includes auxiliary multifunctional brush card machine (7), works as face After identification payment failure, complete to pay by the auxiliary multifunctional brush card machine (7).

3. device according to claim 1 or 2, it is characterised in that: pressure is arranged in the bottom surface of the service plate placement plate (6) Sensor is placed on service plate placement plate (6) with perceiving the vegetable for whether needing to be settled accounts.

4. device according to claim 1 or 2, it is characterised in that: the vegetable identifies camera (5) according to the shadow of input As variation, the vegetable whether automatic sensing needs to be settled accounts is placed on service plate placement plate (6).

5. device according to claim 1 or 2, it is characterised in that: the service plate placement plate (6) is provided with luminous zone, when When vegetable wait settle accounts is placed on service plate placement plate (6), luminous zone is blocked, is to enhance vegetable identification camera (5) perception The no vegetable for needing to be settled accounts is placed in the accuracy rate on service plate placement plate (6).

6. device according to claim 1 or 2, it is characterised in that: device further includes customer service touch screen (2), the customer service Touch screen (2) and customer's touch screen (3) are opposite to be staggered on the cashier (100).

7. device according to claim 1 or 2, it is characterised in that: device further includes voice announcer, and from trend, customer is broadcast Voice prompting is reported, voice prompting content includes whether vegetable detail, dish valence, clearing total amount and clearing to be paid succeed.

8. device according to claim 6, it is characterised in that: the customer service touch screen (2) and customer's touch screen (3) use Liquid crystal touch screen, vegetable identification secondary light source (9) use LED light source.

9. a kind of valuation and cash register control system of view-based access control model identification, which is characterized in that control system includes: vegetable typing mould Block, Customer Information recording module and clearing payment module,

The vegetable recording module takes pictures to the new vegetable in standard service plate (8) by vegetable identification camera (5), will acquire dish Product photo sends processor to, extracts image feature information to vegetable photo by the processor of host (1) and combines artificial treatment, Determine menu name and price；

The Customer Information recording module identifies that camera (4) take pictures to customer head by customer, the customer that will acquire Photo sends processor to, carries out graphic feature extraction to customer's photo by the processor of host (1) and is shown in customer's touch-control Shield on (3), and combines customer's artificial treatment, typing customer face, identity information and payment accounts information；

The clearing payment module takes pictures to the vegetable in service plate placement plate (6) by vegetable identification camera (5), is valuated And dining information is shown on customer's touch screen (3)；Meanwhile identifying that camera (4) capture customer face photo, warp by customer Processor analysis identification customer simultaneously shows Customer Information on customer's touch screen (3)；On customer's touch screen (3) simultaneously Show that customer confirms that the option of payment is confirmed by customer and completes the payment of brush face.

10. control system according to claim 9, it is characterised in that: system further includes auxiliary payment module, the auxiliary For payment module by assisting multifunctional brush card machine (7) to collect money, the auxiliary payment module includes IC cCredit card payment mode, two Tie up code payment mode or quick flashing payment mode.

11. the valuation and cash method of a kind of identification of view-based access control model, which is characterized in that method includes:

Preparation process, preparation process complete the typing of tableware template, are that standard is carried out in vegetable typing and the vegetable identification in use process It is standby；And Customer Information typing is completed, the face information of customer and account information are bound, are ready for the payment of brush face；

Meal use process is opened, meal use process is opened and includes clearing payment when opening vegetable typing before the meal, opening meal and open postprandial Data check process；Wherein vegetable typing will update and adjust history vegetable and price, while the new vegetable of typing and corresponding valence Lattice；Clearing payment realizes clearing in such a way that the vegetable of vision and vegetable identification combine, and by exempting from close 1:N's The mode of brush face payment is paid；Data are checked, the sales situation of every meal is counted, and sales volume, sale including each dish are total Volume, customer's personal consumption volume, customer's history consumption vegetable and customer's history consume total value；

Vegetable and nutrition data push process, by customer's dining information interior for a period of time, dining ingredient and nutrition data, consumption Information and healthy pantry advisory information send customer to.

12. according to the method for claim 11, it is characterised in that: the tableware template typing of the preparation process includes as follows Step:

(1) power on, opening device；So that bright service plate placement plate (6), cooperation camera ginseng are beaten in vegetable identification secondary light source (9) Number obtains clear photograph；There is flexibly stable adaptability photo to obtain environment in different dining rooms with this；

(2) operator gets out tableware all in dining room, and prepares a certain amount of vegetable；

(3) operator's activation system carries out tableware shooting；

(4) operator takes a tableware, and a certain amount of vegetable is packed into tableware, and tableware is placed on to each position of pallet, And the angle of tableware is adjusted, shoot several photos；Photo is stored in same file position；

(5) (4) operator repeats step to each tableware；

(6) after all tableware pictures are ready to complete, the typing of tableware template and learning program are run, completes tableware typing and study；

(7) the tableware of typing can be used in the normal operation in subsequent dining room；When replacing or increasing new tableware, Then delete should not tableware picture, the picture of the new tableware of retake simultaneously repeats (6) typing and study that step completes new tableware.

13. according to the method for claim 11, it is characterised in that: the Customer Information typing of the preparation process includes as follows Step:

(1) opening device, activation system, selection enter Customer Information recording module；

(2) customer according to system suggestion inputs account information, and account information can be with the cell-phone number of association user, identity card on backstage Number, student number, mailbox, address, and/or work number；

(3) after customer inputs information, whether system check is legal account, and confirmation account whether there is；If not conforming to rule prompt It is required that re-entering；

(4) prompt customer faces customer and identifies camera, and prompting mode includes that voice prompting, text prompt, and/or indicator light mention Show；

(5) starting camera program captures client head photo, prompts customer's new line to nod, left and right shakes the head, and/or blinks, photo After the completion of shooting, be sent into photographic quality program, photographic quality is assessed by assessment algorithm, first determine whether photo whether mould Paste is taken pictures again if fuzzy, is extracted human face characteristic point by Feature Extraction algorithm if photographic quality passes through, is judged people Whether face is complete, whether angle is suitable, undesirable, takes pictures, meets the requirements again, carries out in next step；

(6) remind customer to identify successfully, and the picture of angles multiple in shooting process is stored in the association Pictures location of the customer； The face characteristic that face characteristic extraction algorithm extracts all angles is passed through to the photo of customer, and by these features with the customer's Account information association；

(7) typing customer account password pops up input password window and simultaneously prompts input password, when password reply future pays Special circumstances；Prompt inputs password again and verifies；Not by then re-entering, by then carrying out in next step；

(8), if face is crossed in typing before in the account, judge whether new typing face and original face are same People；If it is pass through；Otherwise report an error false information, and requires connection staff；

(9) it is fully completed, prompts typing success.

14. according to the method for claim 11, it is characterised in that: the vegetable typing for opening meal use process includes as follows Step:

(1) opening device, activation system, selection enter vegetable recording module；It opens vegetable identification secondary light source and plays bright service plate placement Plate cooperates camera parameter, obtains clear photograph；There is flexibly stable adaptability photo to obtain ring in different dining rooms with this Border；

(2) operator gets out a sample of each vegetable of the meal, and each vegetable guarantees this in a tableware A vegetable remains loaded into the tableware；

(3) the vegetable for needing typing is placed into standard service plate by operator, and standard service plate is placed into the knowledge in service plate placement plate Other area；

(4) after confirming correct placement, vegetable recognizer is started；Recognizer calls vegetable identification camera to shoot a photo, Vegetable detection program is sent into after verifying quality；Detect the vegetable and sundries in Programmable detection picture, and by vegetable in picture and miscellaneous Object part is split from picture, is sent into sort program, sundries is rejected；

(5), for the vegetable detected, tableware used in the vegetable is identified first；Extract tableware information, the tableware with typing before Template compares, and matches highest several tablewares；If matched highest tableware meets threshold condition, calculate and second The similarity of a similar tableware；Otherwise it is assumed that this tableware is not present in the tableware template for realizing typing, this tableware is needed Want typing tableware template；If tableware exists in a template, and most like and time similar difference is bigger, then determines the tableware For the similar tableware of top1 in tableware template；Otherwise, it dishes out several possible tablewares, allows operator's artificial selection, to ensure standard Exactness；The above matching process for completing tableware；

(6) the matched vegetable of tableware is completed for confirmation, a text box that can be inputted occur；Operator is clicked by touch screen Text frame can input the title and price of the vegetable；

(7), for the vegetable of each completion price setting, picture is sent into vegetable recognizer, passes through the vegetable feature of precondition Extraction algorithm extracts the characteristic of the vegetable；

(8) operator clicks confirmation vegetable typing, and whether all vegetables of program ver-ify have price, and identical vegetable does not have There is different prices, then passes through；All vegetables are entered into the vegetable database of the meal on the standard service plate；

(9) repeat the above process (3)~(8), complete the typing of all vegetables.

15. according to the method for claim 11, it is characterised in that: the clearing payment for opening meal use process includes as follows Step:

(1) opening device, activation system, selection enter clearing payment module；It opens vegetable identification secondary light source and plays bright service plate placement The corresponding cog region of plate has flexibly stable adaptability photo to obtain environment in different dining rooms with this；

(2) the standard service plate for filling vegetable tableware is placed into service plate placement plate by client, and vegetable identifies that camera shoots standard service plate Photo, and send photo to detection program；

(3) the vegetable and sundries in Programmable detection photo picture are detected, by Recognize algorithm trained in advance by dish in picture Product and sundries part are divided from picture, are sent into sort program, sundries is rejected；

(4), for the vegetable detected, the vegetable frame choosing that will test out is divided；

(5) the vegetable identification range confirmed, for some vegetable, first extraction tableware feature, by tableware feature and meal typing Tableware Characteristic Contrast obtains several tablewares before similarity；If top1 tableware is unsatisfactory for similarity by threshold value, determining should Tableware is illegal；If top1 crosses threshold value, the value of top1-top2 is calculated；If the value of top1-top2 is greater than threshold value, The tableware of top1 has conspicuousness, then only need to identify vegetable in the corresponding tableware of top1；If the value of top1-top2 is less than threshold Value, then the tableware of top1 does not have conspicuousness, corresponding vegetable need to be matched in the corresponding tableware of top1 and top2, to ensure standard Determining and reliability；

(6) after the vegetable identification range of confirmation, vegetable is identified: extraction vegetable feature first, vegetable feature and corresponding tableware range Interior vegetable Characteristic Contrast, obtains vegetable sequencing of similarity；If top1 vegetable is unsatisfactory for similarity by threshold value, determining should The non-typing of vegetable；If top1 crosses threshold value, the value of top1-top2 is calculated；If the value of top1-top2 is greater than threshold value, The vegetable of top1 has conspicuousness, then is determined as the vegetable of top1；If the value of top1-top2 is less than threshold value, the vegetable of top1 Without conspicuousness, the match information of comprehensive dish and the match information of tableware, weighted calculation obtains an integrated value, and takes top1 Vegetable is the vegetable identified；Risk is recorded simultaneously, in case retrospect；

Repeat step (5)~(6), complete the identification of all vegetables, and according to the price of typing, obtain the price of vegetable, calculate It eats out total price；

(8) above-mentioned photo with partitioning scheme and shows on customer's touch screen dutiable value, and shows dining total price, text and/ Or voice prompting customer eats and total price and pays；

After vegetable completes valuation or valuation simultaneously, starting customer identifies camera, obtains and cleans complete customer head and shine Piece extracts human face characteristic point, judges whether face is complete, and whether angle is suitable, improper, retakes, properly then carries out next Step；

(10) the face picture that will meet quality is sent into recognition of face program, extracts face characteristic, and and in input database Face characteristic be compared；Obtain the human face photo of similarity top1 and top2；First determine whether the similarity of top1 is full Foot is by threshold value, if conditions are not met, then determining the customer not in user library；Prompt the non-typing of face；At this time typing face or Pass through the mode payment of auxiliary payment；

If the face of top1 meets threshold value, needing judgement is the risk of this people；The value of top1-top2 is calculated, if should Value is greater than risk threshold value, then determines that top1 has conspicuousness；If the value is less than risk threshold value, determine that top1 has risk Property, need additional verifying means to guarantee accurate fixed, at this point, typing password when customer can input typing face carry out it is secondary Verification completes identity validation, and carries out brush face payment.

16. according to the method for claim 11, it is characterised in that: the postprandial data for opening meal use process check packet Include: the dining consumption of the sales data of the total sales data and single vegetable in dining room in statistics a period of time, each customer is total Valence, dining ingredient and nutrition data.